Comment by bigiain

I wonder if there’s a Fourier transform equivalent that’d let you stack up various inverse square curves to appropriate a linear curve?